libitm fails to compile because -march=zEC12 option is used for assembler ... libtool: compile: /builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/./gcc/xg++ -B/builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/./gcc/ -nostdinc++ -nostdinc++ -I/builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/s390-redhat-linux/libstdc++-v3/include/s390-redhat-linux -I/builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/s390-redhat-linux/libstdc++-v3/include -I/builddir/build/BUILD/gcc-4.8.1-20130717/libstdc++-v3/libsupc++ -I/builddir/build/BUILD/gcc-4.8.1-20130717/libstdc++-v3/include/backward -I/builddir/build/BUILD/gcc-4.8.1-20130717/libstdc++-v3/testsuite/util -L/builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/s390-redhat-linux/libstdc++-v3/src -L/builddir/build/BUILD/gcc-4.8.1-20130717/obj-s390-redhat-linux/s390-redhat-linux/libstdc++-v3/src/.libs -B/usr/s390-redhat-linux/bin/ -B/usr/s390-redhat-linux/lib/ -isystem /usr/s390-redhat-linux/include -isystem /usr/s390-redhat-linux/sys-include -DHAVE_CONFIG_H -I. -I../../../libitm -I../../../libitm/config/linux/s390 -I../../../libitm/config/linux -I../../../libitm/config/s390 -I../../../libitm/config/posix -I../../../libitm/config/generic -I../../../libitm -ftls-model=initial-exec -mzarch -mhtm -Wa,-march=zEC12 -Wall -pthread -Werror -std=gnu++0x -funwind-tables -fno-exceptions -fno-rtti -fabi-version=4 -O2 -g -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches -march=z9-109 -mtune=z10 -D_GNU_SOURCE -MT aatree.lo -MD -MP -MF .deps/aatree.Tpo -c ../../../libitm/aatree.cc -fPIC -DPIC -o .libs/aatree.o Assembler messages: Error: invalid switch -march=zEC12 Error: unrecognized option -march=zEC12 ... full log is available from http://s390.koji.fedoraproject.org/koji/taskinfo?taskID=1135021 Version-Release number of selected component (if applicable): gcc-4.8.1-5.fc20 binutils-2.23.88.0.1-8.fc20
Nick, please see Jakub's response on IRC <jakub> dhorak: sounds like a binutils issue <jakub> dhorak: unlike el7 or f19, the f20 binutils aren't based on hjl's releases and thus are significantly older than what f19 or el7 has <jakub> dhorak: you need to bug binutils maintainers to backport the needed changes
power8 support is missing in f20 too I think, don't know what is the AVX 3.1 state.
Jakub, should I expect any problem when binutils-2.23.52.0.1-10.fc20 will be used to build gcc? Older, but from the Linux branch.
binutils-2.23.88.0.1-12.fc20 should fix this. Cheers Nick
Just to clarify, F20 isn't "significantly older", and we made a conscious decision to get away from the HJ releases and instead base Fedora (and ultimately RHEL) on the upstream binutils releases. We expected some problems from this transition, particularly on the lesser tested platforms. It looks like Nick already took care of this instance. Please feel free to contact either of us if there are additional issues.
This bug appears to have been reported against 'rawhide' during the Fedora 20 development cycle. Changing version to '20'. More information and reason for this action is here: https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora20